Treatment outcome has improved in acute myeloid leukemia (AML), AML. To improve treatment outcomes in AML, there is a compelling need to develop treatments which attack novel cellular targets and which are not affected by known mechanisms of resistance. Irinotecan (CPT-11) is a topoisomerase I-interactive drug which, as a single agent, has shown limited activity against 5-florouracil (5-FU)-refractory colorectal cancer, but, in combination with 5-FU, is very effective against the disease both in preclinical models and in clinical trials. The efficacy of this combination is highly schedule-dependent. We have found that CPT-11 is highly active against multidrug-resistant human leukemia xenografts in vivo. Moreover, preclinical data from our laboratory demonstrate that the schedule-dependent synergistic drug interaction which has been found for CPT-11 and 5-FU also applies to CPT-11 in combination with Ara-C. Based on these preclinical findings, a Phase I clinical protocol was designed and initiated combining CPT-11 with Ara-C in the treatment of refractory AML and chronic myelogenous leukemia in myeloid blast transformation (CML-MBT). The objective of this application is to develop CPT-11 and Ara-C combination chemotherapy as a treatment for myeloid leukemias resistant to current therapies; this approach would then subsequently be applied to untreated AML with a low likelihood of response to current regimens. The specific aims presented are: 1. To define the efficacy of CPT-11 and Ara-C combination chemotherapy in refractory AML and in CML-MBT; 2. To optimize the schedule of CPT-11 and Ara-C combination chemotherapy based on laboratory correlates of efficacy; and 3. To identify correlates of sensitivity and resistance to CPT-11 and Ara-C combination chemotherapy.